[ν΄κ³ νμ] λ΄μ€ μ λͺ©(νκΈ): [ν΄κ³ νμ] μλ¬Έ μ λͺ©(μλ¬Έ): Sakana AI and NVIDIA Introduce TwELL with CUDA Kernels for 20.5% Inference and 21.9% Training Speedup in LLMs μλ¬Έ: Sakana AI and NVIDIA Introduce TwELL with CUDA Kernels for 20.5% Inference and 21.9% Training Speedup in LLMs μμ€: marktechpost MD νμΌ: content/2026-05-12/marktechpost-sakana-ai-and-nvidia-introduce-twell-with-cuda-ker.md